CN114398773A - Coal mine dispatching robot system based on artificial intelligence technology - Google Patents

Abstract

The application provides a coal mine dispatching robot system based on an artificial intelligence technology. The system comprises a scheduling AI brain, a three-dimensional visual intelligent scheduling platform and a mine intelligent scheduling platform, wherein the scheduling AI brain comprises an AI platform module and a scheduling AI knowledge map module: the three-dimensional visual intelligent scheduling platform is used for providing a scheduling twin model and realizing three-dimensional visualization of coal mine production commanding and scheduling; and the mine intelligent scheduling platform is used for realizing the functions of a plurality of service scene modules based on the scheduling algorithm model and the knowledge graph. The working efficiency of coal mine dispatching personnel is greatly improved, and the service burden is reduced.

Description

Coal mine dispatching robot system based on artificial intelligence technology

Technical Field

The application relates to the field of intelligent scheduling of coal mines, in particular to a coal mine scheduling robot system based on an artificial intelligence technology.

Background

The production scheduling is used as the core work of coal mine production management, and plays an important role in the production construction of coal mine enterprises. The traditional production scheduling work is carried out in a mode of manually collecting production information data and carrying out statistics, so that the information transmission speed is low, the work is complicated, the error is large, the coverage area is narrow, and the accident probability is increased due to errors in management easily. The main problems in the current production scheduling are as follows:

(1) the dispatching center adopts a dispatching mode mainly based on a traditional manual mode at present. The dispatcher of the dispatching room mainly records information through a telephone, a pen and a paper table. When scheduling command is needed, the timeliness of scheduling is influenced to a certain extent by looking up paper character data information, so that the scheduling efficiency is low, and the coal mine production efficiency is influenced. (2) The dispatcher records various machine accounts and reports by taking a manual mode as a main mode, and then binds the machine accounts and the reports into a book by month. When the device is needed to be used in an emergency, the search is slow, and time and labor are wasted. (3) The communication among all departments of the coal mine is coordinated through a dispatching center, so that the effective communication among all the departments is influenced, and the effective utilization of information cannot be guaranteed.

With the combination of new generation information technologies such as Artificial Intelligence (AI), 5G communication, cloud computing, virtual reality, and the like, and production scheduling management, the accuracy and work efficiency of scheduling work are effectively improved. At present, related dispatching management personnel need to carry out all-round management and dispatching on aspects such as coal mine production personnel, electromechanical equipment, coal transportation, ventilation and the like. In production scheduling, higher requirements are provided for coal mine scheduling personnel, professional coal mine management knowledge is required, and related knowledge in aspects of enterprise management, communication, statistics and the like is required to be known and mastered.

Disclosure of Invention

The application provides a coal mine dispatching robot system based on artificial intelligence technology, and the integrated intelligent production dispatching of a mine, a coal preparation plant and a loading station is realized based on the artificial intelligence technology. The pressure of traditional manual scheduling is reduced, the intelligent operation level of a coal mine scheduling room is improved, the integrated application of big data, an artificial intelligence technology and coal mine scheduling integration is realized, and the problems in the prior art are effectively solved.

The task of the intelligent dispatching robot system for the coal mine is to analyze and sense the operation condition of a coal production system in real time and assist the dispatcher to carry out related dispatching business. Under the background of big coal data, the intelligent coal mine dispatching robot system mainly depends on data learning and other modes to realize the application of an artificial intelligence technology in the field of coal.

The technical scheme of the application is as follows:

the embodiment of the application provides a colliery dispatching robot system based on artificial intelligence technique, including dispatch AI brain, three-dimensional visual intelligent scheduling platform and the intelligent scheduling platform of mine, wherein, the dispatch AI brain includes AI platform module and dispatch AI knowledge map module:

the AI platform module is used for providing a plurality of scheduling algorithm models, and comprises a data acquisition unit, a data marking unit, a model training unit, a model issuing unit and a model deploying unit which are used for constructing the scheduling algorithm models;

the dispatching AI knowledge graph module is used for providing a plurality of knowledge graphs related to coal mine dispatching operation, and comprises a data source acquisition unit, a knowledge extraction unit, a knowledge fusion unit and a graph storage unit which are used for constructing the plurality of knowledge graphs;

the three-dimensional visual intelligent scheduling platform constructs a scheduling twin model of a virtual space which is mapped with an underground physical space scheduling system through a digital twin technology, and fuses and coordinates the acquired real-time data in the production process of a mine and a physical entity to realize three-dimensional visualization of coal mine production commanding and scheduling;

the mine intelligent scheduling platform is used for realizing the functions of a plurality of service scene modules based on a plurality of scheduling algorithm models and a plurality of knowledge maps provided by the scheduling AI brain and a scheduling twin model provided by the three-dimensional visual intelligent scheduling platform.

Furthermore, the plurality of service scene modules comprise an intelligent query function module, the intelligent query function module comprises a production data query unit, a production data prediction unit and an intelligent question and answer unit, and the production data query unit is used for searching corresponding answers from a database according to query questions and feeding back the answers; the production data prediction unit is used for acquiring production prediction data based on the production data and a preset algorithm model; and the intelligent question-answering unit is used for acquiring corresponding answers from the knowledge graph or the database according to the acquired user questions and feeding back the answers.

Furthermore, the plurality of service scene modules comprise an intelligent report function module, and the intelligent report function module is used for constructing distributed uniform metadata for all databases, and completing the writing process of the same metadata of all databases to realize the synchronous filling of a plurality of reports; the intelligent report function module also comprises a data prediction module which is used for obtaining prediction data based on the relevant data corresponding to the business demand and a preset prediction model according to the business demand and filling the prediction data into a corresponding report.

Furthermore, the plurality of service scene modules comprise an intelligent electronic work order function module, and the intelligent electronic work order function module is used for creating an electronic work order and tracking the flow of the electronic work order according to the acquired work order creating request; the intelligent electronic work order function module also comprises a speed dialing unit which is used for inquiring the contact way of the processor according to the processor corresponding to the electronic work order and automatically notifying.

Further, the plurality of service scene modules include an intelligent voice function module, and the intelligent voice function module is configured to confirm the identity of a sender corresponding to the received voice information of the user side through voiceprint recognition, convert the voice information of the user side into text information through a voice recognition technology, and convert the text information fed back to the user side into voice information through a voice synthesis technology.

Further, the plurality of service scene modules comprise an intelligent telephone directory function module, the intelligent telephone directory function module comprises a basic information management unit, a staff telephone inquiry unit and an intelligent telephone recommendation unit, the basic information management unit is used for managing basic information of the telephone directory, the staff telephone inquiry unit is used for acquiring telephone information of staff from the basic information of the telephone directory according to corresponding information in staff, places, equipment and duty situations, and the intelligent telephone recommendation unit is used for automatically generating a contact list and dialing priority according to various event processing procedures.

Further, the plurality of service scene modules comprise a security situation analysis function module, and the security situation analysis function module is used for acquiring the environmental security data, acquiring the change trend of the data index according to a preset security prediction model, and alarming the security data or the change trend according to the acquired change trend.

Furthermore, the plurality of service scene modules comprise an emergency aid decision function module, and the emergency aid decision function module is used for generating an emergency processing flow according to various input preset rules and task flows in the intelligent electronic work order function; and under the condition of emergency, acquiring a recommended processing scheme from the emergency knowledge graph as decision assistance; the escape system is also used for generating an escape route according to at least one of information of an accident occurrence position, a position of an employee, a vehicle position and an accident category under the condition that a person needs to be removed from the underground in the event of a major safety accident.

Furthermore, the plurality of service scene modules comprise a coal quality prediction and management function module, and the coal quality prediction and management function module is used for acquiring a raw coal quality prediction result according to the acquired raw coal quality data and the raw coal quality prediction model; the system is also used for carrying out corresponding risk early warning prompt according to the obtained raw coal quality prediction result and a preset coal quality risk early warning threshold value; and the method is also used for acquiring a corresponding auxiliary solution according to the acquired raw coal quality prediction result.

Further, the plurality of service scene modules comprise a log traceability analysis module, the log traceability analysis module comprises a log collection unit and a log data mining unit, and the log collection unit is used for acquiring log data generated in the system operation process, processing the log data by using a Flume technology and writing the log data into a log traceability server database; the log data mining unit is used for acquiring high-level data from the log data according to the scheduling service analysis result; and analyzing an APRIORI algorithm through association rules to obtain an analysis result from the high-level data.

The technical scheme provided by the embodiment of the application at least has the following beneficial effects:

according to the invention, through deep analysis of the underground coal mine service scene, aiming at complicated scheduling contents, a coal mine scheduling AI brain and man-machine interaction digital twin model is constructed, and thus three-dimensional visual intelligent scheduling is realized. An intelligent scheduling system with comprehensive and differentiated functions is created by combining algorithms such as artificial intelligence natural language processing, machine vision, knowledge maps and the like, so that the service automation and the intelligence of a scheduling room are realized. The working efficiency of coal mine dispatching personnel is greatly improved, and the service burden is reduced.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the application.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the application and are not to be construed as limiting the application.

Fig. 1 is a general architecture diagram of a coal mine intelligent dispatching AI robot system according to an embodiment of the invention;

FIG. 2 is a flow chart of the construction of a scheduling AI knowledge-graph in accordance with an embodiment of the present invention;

FIG. 3 is a diagram showing the main technical components of an intelligent scheduling robot system according to an embodiment of the present invention;

FIG. 4 is a flow chart of information processing of the intelligent electronic work order function module according to an embodiment of the present invention;

FIG. 5 is a diagram illustrating a scheduling component of the intelligent voice function module according to an embodiment of the present invention;

FIG. 6 is a schematic diagram of the composition of an emergency knowledge-graph according to an embodiment of the present invention;

FIG. 7 is a schematic diagram of a raw coal quality prediction process according to an embodiment of the present invention;

fig. 8 is a data processing flow of the log tracing service according to the embodiment of the present invention.

Detailed Description

In order to make the technical solutions of the present application better understood by those of ordinary skill in the art, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

It should be noted that the terms "first," "second," and the like in this application are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the application described herein are capable of operation in sequences other than those illustrated or otherwise described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

The technical solutions in the embodiments of the present invention are clearly and completely described below with reference to the drawings and the detailed description, and the described embodiments are only a part of the embodiments of the present invention, but not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

The task of the intelligent dispatching robot system for the coal mine is to analyze and sense the operation condition of a coal production system in real time and assist the dispatcher to carry out related dispatching business. Under the background of big coal data, the intelligent coal mine dispatching robot system mainly depends on data learning and other modes to realize the application of an artificial intelligence technology in the field of coal.

The invention aims at the existing practical problems in coal mine dispatching, researches a coal mine dispatching robot system based on an artificial intelligence technology, and realizes mine integrated production dispatching. The pressure of traditional manual scheduling is reduced, the intelligent operation level of a coal mine scheduling room is improved, and the integrated application of big data, an artificial intelligence technology and coal mine scheduling integration is realized.

As shown in fig. 1, the coal mine dispatching robot system based on the artificial intelligence technology provided by the embodiment of the invention includes an AI dispatching brain, a three-dimensional visual intelligent dispatching platform based on the digital twin technology, and a mine intelligent dispatching platform formed on the basis of the three-dimensional visual intelligent dispatching platform.

A first part: an AI brain is dispatched, coal mine intelligent equipment is used as a bottom layer support, and real-time sensing, control, management and analysis decision-making of a production site are realized by using an AI intelligent technology. The system mainly comprises an AI platform, an AI knowledge graph scheduling module and a man-machine cooperation module.

(1) And the AI platform module is used for providing various scheduling algorithm models and realizing one-stop development and management of the various scheduling algorithm models, and has the main functions of data acquisition, data marking, model training, model publishing and model deployment.

(2) And the dispatching AI knowledge map module is used for providing various knowledge maps related to coal mine dispatching operation, extracting related entities in coal mine dispatching operation based on coal mine related texts and identifying the relationship among the entities. The scheduling AI knowledge graph construction flow is shown in fig. 2. The establishment of the coal mine AI knowledge graph aims to establish a knowledge network which can be understood by coal practitioners and computers, so that the knowledge network can be used for functions of intelligent search, intelligent question answering, personalized recommendation and the like, and the intelligent construction of the coal industry is assisted. The information content related to the coal mine field is complicated, and a plurality of text information such as system standards, organization structures, technical documents and the like need to be comprehensively considered in daily production work. Through knowledge extraction, the extraction of entities and entity attributes and the identification of relationships among the entities from various text information of a mining area are realized. However, the extraction result contains a large amount of redundant and error information, and then the knowledge map of the system is constructed for the coal mine through knowledge fusion modes such as entity alignment, entity disambiguation, entity link and the like, the production is guided by knowledge, and the knowledge and the production work are integrated.

(3) The man-machine cooperation module is used for applying the AI technology to the process of the automatic maintenance system, can realize the automatic control of the intelligent scheduling system, and optimizes the work efficiency of system maintenance, thereby ensuring the safe and stable state of the scheduling system in the operation process, simplifying the manual operation steps, gradually realizing the unattended operation, and further improving the overall stability of the system through the technology. The main functions include data synchronization, algorithm management, application management, and service management.

A second part: a three-dimensional visual intelligent scheduling platform based on digital twins constructs a scheduling twins model of a virtual space which is mutually mapped with an underground physical space scheduling system through a digital twins technology, and fuses and coordinates the acquired real-time data in the production process of a mine with a physical entity to realize three-dimensional visualization of coal mine production commanding and scheduling.

Specifically, a human, machine, ring and pipe multi-service flow digital twin model is constructed, information flow and physical entities are fused and cooperated, and intelligent assistance, intelligent management and intelligent decision of coal mine production commanding and dispatching are realized based on real-time sensing, analysis, judgment, decision, control and learning data of mine production. The module mainly comprises:

(1) digital twin body module: a digital twin three-dimensional visualization technology is used for constructing a virtual space scheduling digital twin body which is mapped with an underground physical space scheduling system, and the scheduling requirements of underground workers are fed back in real time through commands such as mode recognition, gesture perception or semantic understanding based on AI, so that intelligent coordination and organic fusion of man-machine-ring-management are realized.

(2) And a virtual-real fusion module: by means of fusion of digital twins and features, integration and decision support of a scheduling twins model are achieved, and model reconstruction capability is improved by means of machine learning, data mining, edge calculation and artificial intelligence technologies. The production process intelligent monitoring, the equipment performance real-time monitoring, the three-dimensional visualization of production scenes and the early warning of disaster accidents can be realized through the three-dimensional visualization intelligent scheduling, and the production efficiency of the visualization scheduling is improved.

And a third part: and the mine intelligent scheduling platform is used for realizing the functions of a plurality of service scene modules and realizing the service automation and the intellectualization of the mine industrial and mining scheduling room based on a plurality of scheduling algorithm models and a plurality of knowledge maps provided by the scheduling AI brain and a scheduling twin model provided by the three-dimensional visual intelligent scheduling platform. The plurality of service scene modules mainly comprise an intelligent query function module, an intelligent report function module, an intelligent electronic work order function module, an intelligent voice function module, an intelligent telephone directory function module, a safety situation analysis function module, an emergency auxiliary decision function module, a coal quality prediction and management function module and a log traceability analysis function module.

(1) The intelligent query module mainly comprises a production data query unit, a production data prediction unit and an intelligent question and answer (web end and telephone end) unit, wherein the production data query unit is used for searching corresponding answers from a database according to query questions and feeding back the answers; the production data prediction unit is used for acquiring production prediction data based on the production data and a preset algorithm model; and the intelligent question-answering unit is used for acquiring corresponding answers from the knowledge graph or the database according to the acquired user questions and feeding back the answers.

The inquiry range of the intelligent inquiry module comprises a knowledge base formed by information contained in functions of an intelligent report form, an intelligent electronic worksheet, an intelligent telephone book and the like, data are input into different functional parts by underground staff, dispatchers and the like according to actual production work, the system divides related authorities according to different positions of mine staff, the related authorities can clearly inquire information in the authorities, for example, mine leaders and dispatchers can check information such as production data (including recording time, information such as recorders and the like) and prediction data, and the mine staff can inquire information such as watchmen, event responsible staff and associated event progress through the system.

(2) The intelligent report module is mainly used for constructing distributed uniform metadata for all databases, and completing the writing process of the same metadata of all databases to realize synchronous filling of a plurality of reports; the intelligent report function module also comprises a data prediction module which is used for obtaining prediction data based on the relevant data corresponding to the business demand and a preset prediction model according to the business demand and filling the prediction data into a corresponding report.

The method specifically adopts a mode based on one metadata and one central processing unit to solve the tedious work of daily data filling and report summarizing, and data statistical analysis of a dispatcher. The system mainly comprises an intelligent report form web processing unit, a pdf file automatic generation and printing unit, a data path, a data prediction and auxiliary decision suggestion unit.

(3) The intelligent electronic work order module is mainly used for creating an electronic work order and tracking the flow of the electronic work order according to the acquired work order creating request; the intelligent electronic work order function module also comprises a speed dialing unit which is used for inquiring the contact way of the processor according to the processor corresponding to the electronic work order and automatically notifying.

The intelligent electronic work order module is based on a B/S framework, a front-line worker is provided with an intelligent terminal using and scheduling system, a dispatcher and a leader report the work order through a web client, and the system (server) creates, schedules and manages the work order. The system mainly comprises a multi-type electronic work order flow creating and maintaining unit, an electronic work order automatic supervising unit, an authority management unit, a work order voice recognition unit, a client quick dialing unit and a work order multi-dimensional classification and result tracking unit.

(4) The intelligent voice module is mainly used for confirming the identity of a sender corresponding to the received voice information of the user side through voiceprint recognition, converting the voice information of the user side into text information through a voice recognition technology, and converting the text information fed back to the user side into the voice information through a voice synthesis technology. A more flexible mode is provided for scheduling service intellectualization, a more friendly and more convenient man-machine interaction means is provided, so that the operation complexity of a scheduler is reduced, two hands of the scheduler are liberated, and the working efficiency is improved. The method mainly comprises voiceprint recognition, voice recognition and voice synthesis, can automatically recognize the identity of a speaker, replaces two hands of a dispatcher to make a call, can automatically answer some common inquired questions through voice synthesis, and provides voice service for each service function module.

(5) The intelligent telephone book module mainly comprises a basic information management unit, an employee telephone inquiry unit and an intelligent telephone recommendation unit, wherein the basic information management unit is used for managing basic information of a telephone book, the employee telephone inquiry unit is used for acquiring the telephone information of an employee from the basic information of the telephone book according to corresponding information in the employee, place, equipment and duty condition, and the intelligent telephone recommendation unit is used for automatically generating a contact list and dialing priority according to various event processing procedures. The intelligent scheduling robot system is a basic function of an intelligent scheduling robot system, provides technical support for functions of intelligent query robots, intelligent electronic work orders and the like, can query information and telephone of related personnel in multiple angles according to staff, places, equipment and duty situations, can automatically generate contact lists and call priorities according to various event processing flows, and assists a dispatcher in scheduling.

(6) And the safety situation analysis module is mainly used for acquiring the environment safety data, acquiring the change trend of the data index according to a preset safety prediction model, and alarming the safety data or the change trend according to the acquired change trend.

The method specifically comprises three parts of environment safety, equipment safety and safety setting. The environmental safety also comprises environmental data monitoring and environmental safety analysis, and the environmental safety data monitors harmful gas, temperature, humidity, wind current and mine earthquake data and is displayed by a line graph; the environmental safety analysis comprises three parts, namely model management, safety prediction and safety alarm; the model management part is used for establishing and maintaining the single index and the comprehensive index model; the safety prediction can use a model to independently predict the variation trend of a single index, also can comprehensively predict the safety of the whole environment through multiple indexes, and graphically display the prediction data; the safety alarm comprises an alarm information list of real-time safety monitoring and model safety prediction prompting, and can be screened according to the alarm safety level, the alarm type and the safety responsible person.

The classification and basis of the safety alarm levels, the types of the safety alarms and the historical experience input corresponding to the alarms of all types can be managed through safety setting.

(7) The emergency auxiliary decision-making module is mainly used for generating an emergency processing flow according to various input preset rules and task flows in the intelligent electronic work order function; and under the condition of emergency, acquiring a recommended processing scheme from the emergency knowledge graph as decision assistance; the escape system is also used for generating an escape route according to at least one of information of an accident occurrence position, a position of an employee, a vehicle position and an accident category under the condition that a person needs to be removed from the underground in the event of a major safety accident.

And generating an emergency treatment flow according to various input preset rules and a task flow in the intelligent electronic work order function, displaying a recommended treatment scheme and a recommended treatment flow by the system when an emergency occurs, providing information such as treatment processes, loss conditions, remarks and the like of similar events, and assisting a dispatcher in making an emergency decision. Meanwhile, when major safety accidents happen and people need to be removed from the underground, the function supports the escape route planning of underground staff, and the escape route is generated according to information such as the accident occurrence position, the position of the staff, the vehicle position, the accident category and the like, so that the underground staff are assisted to safely and orderly withdraw the accident site. The emergency management system mainly comprises an emergency knowledge management unit, an emergency decision unit and an escape route planning unit.

(8) The coal quality prediction and management module is mainly used for obtaining a raw coal quality prediction result according to the obtained raw coal quality data and the raw coal quality prediction model; the system is also used for carrying out corresponding risk early warning prompt according to the obtained raw coal quality prediction result and a preset coal quality risk early warning threshold value; and the method is also used for acquiring a corresponding auxiliary solution according to the acquired raw coal quality prediction result.

The method mainly predicts the coal quality of raw coal and generates an auxiliary solution after the coal quality is early warned. The method achieves the aims of 'prediction and forecast in advance, process control in advance and improvement after conclusion'. The method mainly comprises a raw coal quality prediction unit, a coal quality risk early warning unit and an auxiliary solution automatic generation unit.

(9) The log source tracing analysis module mainly comprises a log collection unit and a log data mining unit, wherein the log collection unit is used for acquiring log data generated in the system operation process, processing the log data by using a flash technology and writing the log data into a log source tracing server database; the log data mining unit is used for acquiring high-level data from the log data according to the scheduling service analysis result; and analyzing an APRIORI algorithm through association rules to obtain an analysis result from the high-level data.

The log collection unit comprises two functional modules of log collection and transmission and log recording. The log data mining unit comprises three functional modules of scheduling process standardized analysis, associated data analysis, event query and pipeline.

The data generated by the system operation can be stored in a structured mode, the potential value behind the data can be mined, dynamic monitoring and allocation of resources of each service system can be guaranteed, the relation between equipment faults and environmental factors can be mined on a higher level, and continuous, efficient and stable operation of platform services can be guaranteed.

As shown in fig. 2, the intelligent dispatching robot system provided by the invention constructs a knowledge graph, and fig. 2 embodies data sources and a combing process of entities, attributes and relationships of the knowledge graph. And the understanding of the robot system on the scheduling relation is realized through the knowledge graph.

As shown in fig. 3, the system of the present invention implements language identification and conversion recording by processing natural language in the coal mine scheduling process; and reasonable backup and scheduling of data in the system are performed by adopting a proper data storage mode.

Aiming at the service scene requirements, the intelligent scheduling of the mine service is realized by establishing a knowledge graph related to the scheduling service and a scheduling model based on artificial intelligence, and the following further explains a plurality of service function modules constructed aiming at the main service scene as follows:

the intelligent query function module mainly comprises three units of production data query, production data prediction and intelligent question answering.

(1) Production data is input into the production data query unit through a handheld terminal after underground staff carry out production work, the production data is stored in the SQL database and is used for being queried and called by each system, the production data is directly uploaded underground, work efficiency is improved, and meanwhile, a dispatcher is prevented from being informed by a telephone to cause mistakes and omissions in the information transmission process.

Mine leaders and dispatchers log in a system (account number, face, fingerprint and the like for verification) and can check production data of the current day, the current week and the current month, wherein the production data comprises but is not limited to production time of each shift, overhaul time, shutdown time, propulsion distance, raw coal yield and other information; the production team leader logs in the system (account number, face, fingerprint and the like for verification) and can check the production data of the team on the day, the week and the month, including but not limited to the production time of each shift, overhaul time, shutdown time, propulsion distance, raw coal yield and the like.

Aiming at the condition that errors are possibly generated during information entry, underground staff can apply for modifying the production data entered by the underground staff, and the production data is automatically corrected after the dispatcher agrees; after the dispatcher logs in the system (account number, face, fingerprint and the like are verified), the production data of the current day, the current week and the current month, including but not limited to information such as advancing distance of each shift, raw coal yield and the like, can be viewed.

(2) And the production data prediction unit is divided into real-time production data prediction and future production data prediction.

And (3) predicting real-time production data, namely, estimating the production data of the current time point in proportion by the system according to the last updated production data and the time of the change of the query time and distance data, and triggering the function during query to quickly calculate the result.

And predicting future production data, wherein the system predicts the production data on the current day, the current week and the current month according to the production data, current production data, equipment maintenance arrangement and other information recorded by the underground staff, and the function is triggered at zero point every day for prediction. Mine leaders and dispatchers log in a system (account number, face, fingerprint and the like are verified), real-time data can be checked, and the system can quickly calculate the production data at the moment in real time; the predicted production data of the current day, the current week and the current month can also be viewed, and the information comprises but is not limited to the advancing distance of each shift, the raw coal yield and the like.

(3) The intelligent question-answering unit supports two query services of characters and voice of a Web end and a telephone end, when a mining staff inquires at a system end, the mining staff can input questions in two modes of keyboard input and voice input, voice information is automatically converted into texts through a voice recognition function, answers of related questions are found by entering corresponding databases and knowledge maps according to the questions, and the answers of the questions are fed back to the inquirers in two modes of voice and text. The query content of the intelligent question and answer unit comprises three aspects of production data, event flow and personnel information.

The problem structure and content are identified through semantic analysis, information extraction and other methods, corresponding databases and knowledge maps are entered, and features beneficial to finding answers are extracted from the problems and used as representation of the problems. Training a corresponding classifier model through question-answer pairs in advance, finding a question-candidate answer pair with the highest probability when selecting an answer, wherein the candidate answer is the answer of the query question, and feeding the answer back to a querier.

The intelligent report function module solves the tedious work of daily data filling, report summarizing and data statistical analysis of a dispatcher by adopting a mode based on one metadata and one central processing unit.

Firstly, a Web front end is created based on HTML5, and the system has high performance, cross-platform and Excel high compatibility, brings a close and easy-to-use experience for a dispatcher, and meets the service scenes of Web Excel component development, form document collaborative editing, data filling, Excel report design and the like. The method comprises the steps of constructing metadata, eliminating a data island, firstly carrying out adaptation, and opening the writing process of relevant data of other databases to achieve the synchronous filling function of an intelligent report. And in the second step, all metadata are gradually shared to construct distributed unified metadata.

And the data prediction module presets a plurality of common prediction models including a yield prediction model and an equipment maintenance duration prediction model according to experience, automatically generates various reports, and can provide simple auxiliary decision-making suggestions after filling statistical data for analysis.

The intelligent electronic work order function module respectively transmits the messages uploaded and issued by the tasks aiming at the situation that the work orders are transmitted to the dispatching center from one line and are dispatched to the other line, and the information processing flow of the intelligent electronic work order is shown in figure 4. And (3) reporting a task plan by a worker at the working front line through a terminal (client), triggering a fixed work order task, understanding and filing information through voice recognition in the process, executing a specified script according to a specific link of the task, and finally completing the task.

The method can be divided into reporting and issuing according to the transmission route of the work order task. The module can automatically realize the following steps according to the recorded events and links: reporting the starting and stopping of related affairs of the tunneling work and the fully mechanized mining work; according to seven conventional links of hanging a net and moving a house to a reverse side, personnel contact and automatic supervision are carried out; carrying out maintenance entry and reminding on planned maintenance; for the unscheduled maintenance, generating maintenance information to be input into the monitoring system for a dispatcher according to the system knowledge map region, the responsible person corresponding to the service contact and the event notification; carrying out automatic flow reporting and informing processes of the information of the emergency; intelligent catering is carried out according to the situation of the staff in the current daily meal in the mine, and the like.

The work order distribution function is wholly classified and displayed according to production tasks, safety tasks and comprehensive task division, and after each type of work order events are generated, the processing flow sequentially passes through: and reporting the work order, executing the work order flow, supervising the work order, reporting the work order result, and filing the whole process record.

The intelligent voice function module mainly includes voiceprint recognition, voice recognition and voice synthesis, as shown in fig. 5, the intelligent voice function module can automatically recognize the identity of a speaker, replace two hands of a dispatcher to make a call, and automatically answer some common questions of inquiry through voice synthesis to provide voice service for each service function module.

(1) The voiceprint recognition process firstly extracts voice characteristics, then puts the characteristics into a model for training, and finally searches the result with the highest score or the closest score. In the voice library of the workers in the mining area and the intelligent dispatching robot system, voiceprint recognition can be used for authentication, namely speaker confirmation, and can also be used for call log recording and used for identifying the specific identity of the call object.

(2) The speech recognition interaction model includes a hierarchical language model, which is often implemented by an n-gram statistical language model algorithm, and an acoustic model, which is typically implemented using a time-series sensitive Hidden Markov Model (HMM). The dispatch room personnel typically use Mandarin for dispatch command issuance, so it is sufficient to actively issue Mandarin speech recognition. But base miners and other mining personnel are more accustomed to using local dialects for telephone consultation, which raises the necessity for identification.

(3) The speech synthesis process consists of several parts, the first step is text to phoneme conversion, i.e. converting Chinese characters to pinyin. And the second step is audio segmentation, and each phoneme is obtained and corresponds to a starting point and an end point in the audio. And thirdly, predicting the phoneme duration, and obtaining the phoneme duration, the probability of pronouncing or not and the fundamental frequency by using fundamental frequency prediction. The fourth step is to convert the previous high-level features into an acoustic waveform, i.e., a vocoder.

And the intelligent telephone book function module is used for inquiring the knowledge map according to the information of the event type, the occurrence place, the occurrence equipment and the like, matching to a proper scheme, acquiring the contact persons of each process according to the event processing process and the relationship of staff and equipment, and automatically generating an event contact person list (comprising equipment maintenance personnel, manufacturer contact personnel, nearby production captain, on-duty mine leader, mine leader and the like) and dialing priority for reference of scheduling personnel.

The emergency aid decision function module comprises three units of emergency knowledge management, emergency decision and escape route planning.

(1) An emergency knowledge graph, as shown in fig. 6, contains information about events, plans, procedures, cases, personnel, equipment, etc. The event refers to accident types such as fire, power failure and the like, the scheme refers to different processing schemes of different events due to site conditions, the flow refers to each step of specific implementation in the scheme, the case refers to the past accident record of a mining area, and personnel and equipment are basic information of the mining area. According to the safety accident reporting records, the emergency assistant decision-making module stores accident records (including information such as incident time, incident places, accident types, accident causes, processing processes, loss conditions, remarks and the like) as case information into a knowledge graph for knowledge mining, associates the cases with information related to equipment and the like, and associates the cases with processing schemes by dispatchers.

(2) And the emergency decision unit is used for inquiring equipment (including equipment with an accident, other equipment on the upper level and the lower level of the equipment, other equipment near the accident site, common accident associated equipment and the like) associated with the emergency event through the emergency knowledge base according to the environment monitoring data, calling a safety situation analysis function, acquiring the equipment and environment data and assisting a dispatcher to know accident site information. According to report information of underground staff or information input by a dispatcher, the event type is identified, the event content is matched with the accident cause of the case contained in the event type in the knowledge graph, and the detailed content (containing information such as incident time, incident place, accident cause, processing process, loss condition, remark and the like) of the case with the highest similarity and the scheme flow are presented to the dispatcher for reference by the dispatcher. The dispatcher can dispatch by himself or select a reference scheme, and the corresponding intelligent electronic work order function is started. The dispatcher can check the progress of the emergency flow, the gray frame shows the completed flow, the red frame shows the current ongoing flow, and the information such as the conversation content and the feedback condition of workers in the mining area in the flow is displayed, so that support is provided for the dispatcher to master the whole flow, and the emergency event processing efficiency is improved.

(3) The escape route planning unit judges whether people need to be removed or not through two aspects: firstly, judging whether major disaster accidents happen or not through an early warning system with a safety situation analysis function; and secondly, after the report content of underground personnel is detected and the conversation content is synchronously converted into characters, whether the report content accords with major disaster accident signs (such as water inrush signs such as sudden increase of water inflow, fire signs such as smelling oil smell and other accident signs) is judged, and a dispatcher is reminded. And the dispatcher judges whether emergency person removal is needed or not according to system reminding and mine leader indication, and triggers a person removal command. When natural disasters caused by underground flood, fire, electromechanical transportation, gas, coal dust, a roof, disastrous weather and other accidents seriously threatening the safe production of a mine occur, the system extracts and constructs an undirected weighted network diagram according to the topological structure of an underground roadway, uses path planning algorithms such as Dijkstra algorithm or ant colony algorithm and the like, generates escape modes and routes according to information such as accident occurrence positions, positions of employees, vehicle positions, accidents and whether the routes are suitable for driving and the like, and assists the underground employees to safely and orderly evacuate accident areas to the ground or refuge chambers.

The coal prediction and management function module is used for realizing effective control of multiple indexes of moisture, ash content, sulfur content, granularity, sundries and the like of coal. The coal prediction and management function module applies an artificial neural network to the actual prediction of coal quality, as shown in fig. 7, in the modeling stage, historical coal quality inspection data, indexes and heating values corresponding to the historical coal quality inspection data are manually input, a standard library is built, and the standard library is sent to an LSTM (Long Short-Term Memory) network for training to form a raw coal quality prediction model. Among them, LSTM is a time-recursive neural network suitable for processing and predicting important events with relatively long intervals and delays in time series. In the application stage, firstly, noise reduction processing is carried out on sampling data by utilizing a five-point triple sliding average method, and then prediction is carried out through a raw coal quality prediction model to obtain a prediction result.

The log source tracing analysis function module comprises a log collection unit and a log data mining unit.

(1) A data processing flow of the log tracing service provided by the log tracing analysis function module is shown in fig. 8, a log tracing function foundation is established based on data centralization, normalization and relationship storage, and a distributed, reliable and highly available log system for collecting, aggregating and transmitting mass logs is created based on Apache flux. Services such as intelligent query, intelligent report forms and electronic work orders can be customized in a log system as a data sender for collecting data, and simultaneously, the data is simply processed by using flash and written into a database of a log traceability server for managing, analyzing and mining.

(2) The log data mining unit firstly needs to carry out standardized analysis on the scheduling process, analyzes a data set of a higher level to obtain suggestions for system potential data association display and concern points under a macroscopic view angle through a higher level data accumulation and data processing of a service log and an association rule analysis APRIORI algorithm, and further utilizes log resources.

The coal mine intelligent dispatching robot system provided by the embodiment of the invention performs data learning by means of data acquisition and analysis, and mainly utilizes technologies such as big data, deep learning, machine learning, mode recognition and natural language processing. The data source mainly relates to real-time data of multiple production links such as fully mechanized mining, fully mechanized excavation and transportation.

According to the invention, through deep analysis of underground coal mine service scenes, aiming at complicated scheduling contents, a coal mine scheduling AI brain and man-machine interaction digital twin model is constructed, and then a three-dimensional visual intelligent scheduling room is designed and established. An intelligent scheduling system with comprehensive and differentiated functions is created by combining algorithms such as artificial intelligence natural language processing, machine vision, knowledge maps and the like, so that the service automation and the intelligence of a scheduling room are realized. The working efficiency of coal mine dispatching personnel is greatly improved, and the service burden is reduced.

It should also be noted that the exemplary embodiments mentioned in this patent describe some methods or systems based on a series of steps or devices. However, the present invention is not limited to the order of the above-described steps, that is, the steps may be performed in the order mentioned in the embodiments, may be performed in an order different from the order in the embodiments, or may be performed simultaneously.

Other embodiments of the present application will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the invention following, in general, the principles of the application and including such departures from the present disclosure as come within known or customary practice within the art to which the invention pertains. It is intended that the specification and examples be considered as exemplary only.

It will be understood that the present application is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the application is limited only by the appended claims.

Claims (10)

1. The coal mine dispatching robot system based on the artificial intelligence technology is characterized by comprising a dispatching AI brain, a three-dimensional visual intelligent dispatching platform and a mine intelligent dispatching platform, wherein the dispatching AI brain comprises an AI platform module and a dispatching AI knowledge map module:

the AI platform module is used for providing a plurality of scheduling algorithm models, and comprises a data acquisition unit, a data marking unit, a model training unit, a model issuing unit and a model deploying unit which are used for constructing the scheduling algorithm models;

the dispatching AI knowledge graph module is used for providing a plurality of knowledge graphs related to coal mine dispatching operation, and comprises a data source acquisition unit, a knowledge extraction unit, a knowledge fusion unit and a graph storage unit which are used for constructing the plurality of knowledge graphs;

the three-dimensional visual intelligent scheduling platform constructs a scheduling twin model of a virtual space which is mapped with an underground physical space scheduling system through a digital twin technology, and fuses and coordinates the acquired real-time data in the production process of a mine and a physical entity to realize three-dimensional visualization of production commanding and scheduling of a coal mine;

the mine intelligent scheduling platform is used for realizing the functions of a plurality of service scene modules based on a plurality of scheduling algorithm models and a plurality of knowledge maps provided by the scheduling AI brain and a scheduling twin model provided by the three-dimensional visual intelligent scheduling platform.

2. The system according to claim 1, wherein the plurality of business scenario modules comprise an intelligent query function module, the intelligent query function module comprises a production data query unit, a production data prediction unit and an intelligent question and answer unit, and the production data query unit is configured to search a database for corresponding answers according to query questions and perform feedback; the production data prediction unit is used for acquiring production prediction data based on the production data and a preset algorithm model; and the intelligent question-answering unit is used for acquiring corresponding answers from the knowledge graph or the database according to the acquired user questions and feeding back the answers.

3. The system according to claim 1, wherein the plurality of business scenario modules comprise an intelligent report function module, and the intelligent report function module is configured to construct distributed unified metadata for all databases, and to get through a writing process of the same metadata of all databases, so as to implement synchronous filling of a plurality of reports; the intelligent report function module also comprises a data prediction module which is used for obtaining prediction data based on the relevant data corresponding to the business demand and a preset prediction model according to the business demand and filling the prediction data into a corresponding report.

4. The system of claim 1, wherein the plurality of business scenario modules comprises an intelligent electronic work order function module, and the intelligent electronic work order function module is configured to create an electronic work order and perform tracking of an electronic work order flow according to the acquired work order creation request; the intelligent electronic work order function module also comprises a speed dialing unit which is used for inquiring the contact way of the processor according to the processor corresponding to the electronic work order and automatically notifying.

5. The system according to claim 1, wherein the plurality of service scenario modules include an intelligent voice function module, and the intelligent voice function module is configured to confirm the identity of the sender corresponding to the received voice information of the user side through voiceprint recognition, convert the voice information of the user side into text information through a voice recognition technology, and convert the text information fed back to the user side into voice information through a voice synthesis technology.

6. The system according to claim 1, wherein the plurality of service scenario modules include an intelligent phone book function module, the intelligent phone book function module includes a basic information management unit, a staff phone query unit, and an intelligent phone recommendation unit, the basic information management unit is configured to manage basic information of a phone book, the staff phone query unit is configured to obtain phone information of a staff from the basic information of the phone book according to corresponding information in the situations of the staff, the location, the equipment, and the duty, and the intelligent phone recommendation unit is configured to automatically generate a contact list and a dialing priority according to various event processing procedures.

7. The system according to claim 1, wherein the plurality of service scenario modules include a security situation analysis function module, and the security situation analysis function module is configured to obtain environmental security data, obtain a change trend of a data index according to a preset security prediction model, and alarm the security data or the change trend according to the obtained change trend.

8. The system according to claim 1, wherein the plurality of business scenario modules include an emergency aid decision function module, and the emergency aid decision function module is configured to generate an emergency processing flow according to each preset rule entered and a task flow in the intelligent electronic work order function; and under the condition of emergency, acquiring a recommended processing scheme from the emergency knowledge graph as decision assistance; the escape system is also used for generating an escape route according to at least one of information of an accident occurrence position, a position of an employee, a vehicle position and an accident category under the condition that a person needs to be removed from the underground in the event of a major safety accident.

9. The system of claim 1, wherein the plurality of service scenario modules include a coal quality prediction and management function module, and the coal quality prediction and management function module is configured to obtain a raw coal quality prediction result according to the obtained raw coal quality data and a raw coal quality prediction model; the system is also used for carrying out corresponding risk early warning prompt according to the obtained raw coal quality prediction result and a preset coal quality risk early warning threshold value; and the method is also used for acquiring a corresponding auxiliary solution according to the acquired raw coal quality prediction result.

10. The system according to claim 1, wherein the plurality of service scenario modules comprise a log traceability analysis module, the log traceability analysis module comprises a log collection unit and a log data mining unit, and the log collection unit is configured to obtain log data generated during the operation of the system, process the log data by using a flash technology, and write the log data into a log traceability server database; the log data mining unit is used for acquiring high-level data from the log data according to the scheduling service analysis result; and analyzing an APRIORI algorithm through association rules to obtain an analysis result from the high-level data.

Images